RISC-V and ARM are also the main processor cores of embedded devices, and the shipment of RISC-V chips is increasing. For example, on November 21, Pingtou launched three Xuantie RISC-V processors: C907, which realizes AI matrix expansion for the first time, and meets Vector1The C920 with the 0 standard, as well as the real-time processor R910, promote the application of RISC-V in autonomous driving, artificial intelligence, enterprise-class SSD, network communications, and other fields. On November 30, Renesas Electronics successfully designed, tested, and launched a 32-bit CPU core based on the open standard RISC-V instruction set architecture (ISA), the industry's first manufacturer to independently develop a CPU core for the 32-bit general-purpose RISC-V market, with a CoreMark MHz performance of the RISC-V CPU reaching 327, exceeding comparable architectures in the industry, including scaling that improves performance while reducing the amount of **. 
Source: Renesas Electronics On the application side, RISC-V and ARM are currently penetrating into high-performance computing fields such as servers and AI applications, what is the future development trend of the two?The following compares the two processor architectures from a price/performance perspective. Licensing and business models determine costsArm has successfully IPO after a failed takeover, but despite the changing market and competitive environment, Arm is trying to expand revenue. Previously, it was reported that ARM plans to change the royalty collection model and adjust the related fees.
Instead of charging chipmakers royalties based on the value of the chip, Arm's proposed new licensing model will set fees based on the average selling price (ASP) of mobile devices.
The object of licensing fees has also changed, and Arm may no longer charge 2% of the patent fee per chip to chip manufacturers such as Qualcomm, MediaTek, and Samsung, but instead charge a license fee of 1-2% of the selling price of terminal products of equipment manufacturers (such as smartphone manufacturers) that are ultimately equipped with these chips. According to the agency's data, smartphone shipments are divided into high-end, mid-range, and low-end, excluding Apple. Assuming royalties of $12, $6 and $3 per phone per type, respectively, incremental revenue and profit would generate an additional $5.4 billion in operating profit for ARM, which is almost 10 times ARM's operating income.
Caption: ARM has proposed a new licensing model, and the legend shows that the licensing fees generated by non-Apple mobile phones will significantly increase ARM's revenue (Source: SemiAnalysis) It should be seen that ARM has a relatively large market share in the field of mobile devices and embedded devices, and data shows that more than 95% of smartphones are directly or indirectly using the ARM-licensed instruction set. Once ARM's new licensing model is implemented, the impact is enormous. The most obvious impact in the field of mobile phones is that downstream OEMs need to pay chip fees to Arm in addition to licensing fees to Qualcomm, which directly causes a significant increase in costs.
Overall, RISC-V has unique advantages in terms of licensing methods, processor design, cost, and industry chain relationships, and companies should consider these advantages when choosing an architecture for their projects.
Caption: Comparison of licensing and business models of RISC-V and ARM (Chipcheck tabulation).Performance Comparison:ARM has the blessing of the ecosystem, and the module customization of RISC-V has prospectsLet's take Cortex-A76 and P670 as examples to compare the performance of RISC-V and ARM architectures. Arm's Cortex-A78 is slightly higher than Sifive's P670 (using RISC-V) in terms of peak single-threaded performance. Although the Cortex-A78 dominates in terms of raw performance, the P670 is twice as dense as it. So, considering that SiFive's P670 processor is half the physical size of its competitors, its peak single-threaded performance is comparable to that of the Cortex-A78. 
Caption: Arm's Cortex-A76 and Sifive's P670 performance (Source: Makeuseof) The Cortex-A78 was launched in December 2020, and the P670 was unveiled on November 1, 2022, with a nearly two-year R&D gap between the two. Running on the ARMV9 ISA, ARM's latest processors are a substantial improvement over ARMV8 from the Cortex-A78, which delivers about 30 percent higher performance and 50 percent more energy efficiency. In terms of pure performance, ARM processors maintain the lead, however SiFive's P670 has twice the compute density of the Cortex-A78, putting the RISC-V processor in a strong position for compact processors, especially in wearables. Energy efficiency ratio is another aspect of performance comparison, and both RISC-V and ARM use the Reduced Instruction Set Computing (RISC) concept, so it is necessary to study the measurement of energy efficiency performance when selecting a technology. ARM's focus on energy efficiency, combined with its broad ecosystem and processor family, enables the development of energy-efficient devices across a wide range of industries and applications. While the RISC-V holds great promise due to its customization potential, its openness requires a more time and resource investment to take full advantage of its energy-saving features. 
Caption: **The business and tools are relatively rich, and the RISC-V ecosystem is gradually becoming stronger (Source: Xinchacha SaaS).Both are equally capable of navigating AI and automotive applicationsRISC-V and ARM have their own advantages in embedded systems and the Internet of Things, but they have their own advantages in high-performance computing and automotive. Customization continues to be RISC-V's magic weapon in the field of high-performance computing, and the modularity of the RISC-V ISA allows the development of high-performance processors with custom extensions and hardware accelerators to efficiently handle complex workloads such as AI, machine Xi, and big data analytics. Examples of RISC-V-based HPC processors include the SIFIVE U74 and the RISC-V BOOM disordered superscalar processors. In addition, RISC-V is being explored for use in safety-critical and real-time systems such as automotive, aerospace, and industrial control applications. The open-source nature of the RISC-V enables transparent and thorough verification of processor designs, which is critical to ensuring the security and reliability of critical systems. ARM processors are also increasingly being adopted by the data center and high-performance computing (HPC) markets. The ARM Neoverse platform, which includes Neoverse N1 and E1 processors for cloud infrastructure and edge computing applications to deliver high performance and power efficiency for data center workloads, has been commercialized, such as the NVIDIA GH200 Grace Hopper superchip with 72 ARMs Neoverse cores, combined with a combination of GPUs from NVIDIA, deliver up to 10x faster AI performance than x86-based systems. 
Caption: The NVIDIA GH200 Grace Hopper superchip uses the Arm Neoverse with NVIDIA GPUs to meet the needs of AI performance (Source: Arm) In addition, ARM processors are also used in safety-critical and real-time systems such as automotive, aerospace and industrial control applications. Optimized for real-time systems, the Cortex-R series offers fast interrupt response times and deterministic behavior, making it suitable for these demanding applications. SummaryRISC-V still needs to be strengthened, such as the fragmentation of the ecosystem and the lack of standardization caused by product and software compatibility issues. However, RISC-V is of great significance to China, as Academician Ni Guangnan mentioned in recent industry activities, RISC-V has been released for more than ten years, and has gone through the achievements that other chips may take decades to achieve in the past, and 50% of RISC-V's cumulative shipments come from China. At present, China is vigorously developing automotive electronics and AI applications, and RISC-V can seize these opportunities to accelerate development.